Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - JeGX

Pages: 1 ... 18 19 [20] 21 22 ... 37
After some hacking and tweaking on a particle simulator (I tool the original code from the OpenGL Programming Guide 8th Edition), I ended up with this unwanted particle painter. Particles are stored on a SSBO and processed by a compute shader.

GeeXLab - english forum / GLSL Hacker
« on: June 04, 2014, 04:32:46 PM »
A new dev version of GLSL Hacker is ready for Windows 64-bit.

You can download it from this page:

Version - 2014.06.04
! updated the OpenGL context creation under Windows and Linux for when
  the required major version is 2.
+ added create_ex_v3() to gh_render_target lib.
+ added uniform_camera_matrices() and uniform_object_matrix() to gh_gpu_program lib.

GLSL Hacker adds new functions used in my deferred rendering tests.

GeeXLab - english forum / (WIP) Deferred rendering test
« on: May 29, 2014, 10:30:22 PM »
I took some time this afternoon to play a bit with deferred rendering. The last screenshot shows 200 dynamic lights.

MSI, world leading in motherboard technology, is pleased to announce its unleashing of the Z97 GAMING 9 AC motherboard to the public. The Z97 GAMING 9 AC from MSI is a custom designed motherboard, decorated with luscious new heat-sinks and unique Dragon Armor that highlight an arsenal of new features found on the board. It's masterpiece being Xtreme Audio DAC, a premium dedicated onboard sound card with specially selected high quality audio components, set up and tuned by professionals delivering 120dB Signal-to-Noise-Ratio (SNR) 192kHz / 32-Bit of purest sound quality for studio-grade Hi-Fi solutions and gaming headgear. The MSI Z97 GAMING 9 AC also comes with an Intel® Wi-Fi AC module with Bluetooth 4.0 and Intel® WiDi support.

Press release:

The GTX 780 model has double memory: 6GB, while both have 10-phase power design that reduces power noise by 30% and increases efficiency by 15% versus reference designs. The cards are built with DIGI+ VRM and hard-wearing POSCAPs, while the DirectCU-like cooler has been upgraded with ’0dB technology’. When the GPU temp is below 65C the card will run in passive mode with the fans turned off. The large heatsink works to efficiently absorb and dissipate the minimal heat generated by low loads, so while you’re surfing the internet, watching videos, playing light games – your graphics card won’t make a sound.


GeeXLab - english forum / GLSL Hacker
« on: May 28, 2014, 02:45:59 PM »
A new dev version of GLSL Hacker is ready but only for Windows.

You can download it from this page:

Version - 2014.05.28
* fixed a bug in gh_mesh.set_vertices_color() in Lua.
+ added set_vertices_color() to gh_mesh in Python.
+ added mouse_get_wheel_delta() to gh_input lib.
+ added new keyboard keys support (shift, ctrl, alt and windows key).
+ added keyboard_codes.lua in GLSL Hacker Lua libs folder.
+ added FreeImage plugin with floating point (96 and 128-bit) support.
+ added set_current_image_codec_name() to gh_texture module.

GLSL Hacker adds a new image loader plugin based on FreeImage. To use this plugin, just specify it before loading a texture:

Code: [Select]

tex = gh_texture.create_from_file(...)

If the FreeImage plugin is not present, GLSL Hacker uses the built-in image loader.

GeeXLab - english forum / GLSL Transitions
« on: May 24, 2014, 06:16:49 PM »
The  code sample pack has been updated with some cool GLSL transitions intrroduced in this article:

GLSL transitions demos are available in the host_api/GLSL_Transitions/ folder of the code sample pack.


Vendor A has “the most capable GL devs in the industry and the best testing process.” It focuses on OpenGL implementations that work rather than on spec purity. It’s sexy. It’s fun. It’s also the vendor that’s pushing OpenGL as an alternative to Mantle or D3D12.

...this vendor (vendor A) will do things like internally replace entire shaders for key titles to make them perform better (sometimes much better). Most drivers probably do stuff like this occasionally, but this vendor will stop at nothing for performance.

Next up, Vendor B. Vendor B knows the OpenGL spec inside and out and focuses on implementing OpenGL correctly. Unfortunately, Vendor B is characterized as being “a complete hodgepodge, inconsistent performance, very buggy, inconsistent regression testing [and] dysfunctional driver threading…

And Vendor C? Vendor C has two major driver development teams, but doesn’t focus on graphics and is years behind Vendors A and B in implementing basic OpenGL features. It leads the way in open source graphics drivers, and has (mostly) public hardware specs. One driver is years behind but mostly stable as long as you don’t care about performance or features, while the other team’s driver is “a complete disaster…

Vendor A = NVIDIA
Vendor B = AMD
Vendor C = Intel

Complete story:


GeeXLab - forum en français / Filtre de post processing: Toonify
« on: May 24, 2014, 11:03:25 AM »
Voici un petit filtre GLSL sympa qui permet de toonifier ou cartooniser une image:



GeeXLab - english forum / GLSL Hacker
« on: May 21, 2014, 12:43:20 PM »
A new dev version of GLSL Hacker is ready but only for Windows and Max OS X.

You can download it from this page:

Version - 2014.05.19
+ added emitter_set_spawn_positions() to gh_ps module.
+ added back_face_culling() and front_face_culling() to gh_renderer.
+ added box_create(), box_update_size() and sphere_create() to gh_utils.
+ added update_bounding_volume(), get_bounding_box_size(), get_bounding_sphere_center(),
  get_bounding_sphere_radius(), get_global_bounding_box_size(), get_global_bounding_sphere_center(),
  get_global_bounding_sphere_radius() to gh_object.

I added in the Windows version the PolyGon 3D object viewer in the apps/Polygon3DViewer/ folder.  Launch app_start_polygon_3d_object_viewer.bat to start GLSL Hacker with the Polygon3DViewer. Now you can drop any *.OBJ, *.3DS or *.FBX file.

GeeXLab - english forum / (WIP) PhysX fluids - overflowing
« on: May 19, 2014, 09:15:58 PM »
A test with PhysX particles (fluid) and collisions with a mesh (a container). The overflowing of fluid is perfectly handled.

GeeXLab - english forum / (WIP) - PolyGon 3D Object Viewer
« on: May 17, 2014, 11:33:16 PM »
Here are the first screenshots of PolyGon, a 3D object viewer coded with GLSL Hacker.

Under Windows, just drag and drop any 3D object file (*.3ds, *.obj, *.fbx) and is displayed with some basic info.
Drag and drop is managed by DRAG_N_DROP scripts (currently under Windows only).

GeeXLab - english forum / GeForce GTX LED Logo Demo
« on: May 16, 2014, 05:17:55 PM »
There is a new demo in the code sample pack that shows how to control the LED logo of recent GeForce GTX cards.

The demo is available in the host_api/GEFORCE_GTX_LED/ folder.

The GEFORCE GTX LED is managed by three functions in the gh_gml module:

- gh_gml.geforce_logo_is_illumination_supported()
- gh_gml.geforce_logo_get_illumination()
- gh_gml.geforce_logo_set_illumination()

Here is a code snippet in Lua that checks the support of LED logo and sets the illumination to the max (min=0,  max=100):

Code: [Select]
gpu_index = 0  -- first GPU
if (gh_gml.geforce_logo_is_illumination_supported(gpu_index) == 1) then
  gh_gml.geforce_logo_set_illumination(gpu_index, 100)

GeeXLab - english forum / Mesh PQ Torus Knot demo
« on: May 16, 2014, 03:21:24 PM »
GLSL Hacker adds the creation of PQ torus knot. A demo is available in the host_api/Mesh_PQTorusKnot/ folder of the code sample pack.

A gallery of various PQ torus knots is available here:

The following code:
Code: [Select]
steps = 1024
facets = 32
scale = 4.0
thickness = 0.5
clumps = 0
clump_offset = 0
clump_scale = 0.0
u_scale = 2
v_scale = 32
P = 3
Q = -7
pqtorus = gh_mesh.create_pqtorus(steps, facets, scale, thickness, clumps, clump_offset, clump_scale, u_scale, v_scale, P, Q)

will create this torus knot:

GeeXLab - english forum / GLSL Hacker
« on: May 16, 2014, 02:44:34 PM »
A new dev version of GLSL Hacker is ready but only for Windows and Max OS X.

You can download it from this page:

Version - 2014.05.15
+ added OpenGL 4.0 shader subroutines support (gh_gpu_program.set_uniform_subroutine()).
+ added get_capability_4i() to gh_renderer.
+ added create_pqtorus() to gh_mesh.
+ added actor_add_force(), actor_add_force_at_position() and actor_set_kinematic_target()
  to gh_physx3 lib (lua/python).
* fixed a bug in AntTweakBar functions when the AntTweakBar plugin is not loaded.
+ added set_simulation_scales() to gh_physx3 lib (lua/python).

GeeXLab - english forum / GLSL compiler: NVIDIA vs AMD vs Intel
« on: May 15, 2014, 03:13:05 PM »
Here is a thread where I store issues I encounter with NVIDIA, AMD and Intel GLSL compilers.

The following GLSL line is accepted by NVIDIA compiler (R337.50) but generates an error with Intel and AMD compilers:
Code: [Select]
  vec3 tc = vec4(0);

Of course the right line is
Code: [Select]
  vec3 tc = vec3(0);

Pages: 1 ... 18 19 [20] 21 22 ... 37